Overture API Documentation API Reference

The Overture models API allows you to integrate cutting-edge computer vision models into your own applications. The API is completely REpresentational State Transfer (REST) based and allows you to interact with your custom-built deployed models. Please note that the platform is still in beta phase. Using the REST API for production use cases is entirely on your own responsibility. We are happy to receive any positive and constructive feedback on [email protected]. In order to use this REST API you will need an API key, log in on the platform to get your personal key.

API Endpoint
https://models.overture.ai
Terms of Service: https://overture.ai/terms/
Contact: [email protected]
Schemes: https
Version: 1.0.0

Authentication

api_key

The Overture platform uses API keys to authorize API calls. Every model that is deployed is associated with its own API key. API Keys can be retrieved through https://app.overture.ai/deployments

type
apiKey
name
x-api-key
in
header

Image Classification

REST API for image classification.

POST /predict/image/classification

Perform a prediction on images using your own custom-trained image classification model

JSON object containing a list of Base64 URL-safe encoded images.

Request Content-Types: application/json
Request Example
{
  "user_id": "string",
  "model_id": "string",
  "images": [
    {
      "filename": "string",
      "image": "string"
    }
  ]
}

Successful analysis

400 Bad Request

Input Error

401 Unauthorized

Authentication Error

Response Content-Types: application/json
Response Example (200 OK)
[
  {
    "filename": "string",
    "confidence": {
      "label_1": "float",
      "label_2": "float",
      "label_n": "float"
    }
  }
]
Response Example (400 Bad Request)
{
  "error": "string"
}
Response Example (401 Unauthorized)
{
  "error": "string"
}

Face Recognition

REST API for face recognition.

POST /predict/image/facerecognition

Perform a prediction on images using your own custom-trained face recognition model

JSON object containing a list of Base64 URL-safe encoded images.

Request Content-Types: application/json
Request Example
{
  "user_id": "string",
  "model_id": "string",
  "images": [
    {
      "filename": "string",
      "image": "string"
    }
  ]
}

Successful analysis

400 Bad Request

Input Error

401 Unauthorized

Authentication Error

Response Content-Types: application/json
Response Example (200 OK)
[
  {
    "input": "string",
    "info": {
      "width": "integer",
      "height": "integer"
    },
    "output": [
      {
        "person": "string",
        "bounding_coordinates": [
          "integer"
        ]
      }
    ]
  }
]
Response Example (400 Bad Request)
{
  "error": "string"
}
Response Example (401 Unauthorized)
{
  "error": "string"
}

Object Detection

REST API for object detection.

Health

Check the health of your custom trained AI model

POST /health/{user_id}/{model_id}

Check if a certain model is up and running.

user_id: object
in path

10 char long user identifier

model_id: object
in path

integer identifier describing which model to target

200 OK

Model is up and running

401 Unauthorized

Authentication Error

Response Content-Types: application/json
Response Example (401 Unauthorized)
{
  "error": "string"
}

Schema Definitions

ImageInput: object

Image input cannot have more than 15 images for image classification, 5 images for object detection and 10 images for face recognition.

user_id: string
model_id: string
images: object[]
object
filename: string
image: string
Example
{
  "user_id": "string",
  "model_id": "string",
  "images": [
    {
      "filename": "string",
      "image": "string"
    }
  ]
}

ImageClassificationResult: object

filename: string

the filename of the uploaded image

confidence: object

The confidence score per label

label_1: float 0 ≤ x ≤ 1

The confidence score for 'label_1'

label_2: float 0 ≤ x ≤ 1

The confidence score for 'label_2'

label_n: float 0 ≤ x ≤ 1

The confidence score for 'label_n'

Example
{
  "filename": "string",
  "confidence": {
    "label_1": "float",
    "label_2": "float",
    "label_n": "float"
  }
}

FaceRecognitionResult: object

input: string

filename of the uploaded file

info: object
width: integer

width of the uploaded picture

height: integer

height of the uploaded picture

output: object[]
object
person: string

label of the person that has been detected on the image

bounding_coordinates: integer[]
integer

An array of 4 integers that represents a bounding box. Array[0] contains the X-coordinate, array[1] contains the Y-coordinate, array[2] contains the width of the bounding box and array[3] contains the height of the bounding box.

Example
{
  "input": "string",
  "info": {
    "width": "integer",
    "height": "integer"
  },
  "output": [
    {
      "person": "string",
      "bounding_coordinates": [
        "integer"
      ]
    }
  ]
}

ApiError: object

error: string INVALID_LIST_IMAGES_UPLOADED, INVALID_MODEL_ID, INVALID_USER_ID
Example
{
  "error": "string"
}

AuthenticationError: object

error: string NO_VALID_AUTHENTICATION_PROVIDED
Example
{
  "error": "string"
}

ApiResponse: object

code: integer (int32)
type: string
message: string
Example
{
  "code": "integer (int32)",
  "type": "string",
  "message": "string"
}